Journal of Pharmaceutical and Biomedical Analysis 107 (2015) 386–393
Contents lists available at ScienceDirect
Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba
High-performance liquid chromatography method for the determination of hydrogen peroxide present or released in teeth bleaching kits and hair cosmetic products Pascal Gimeno ∗ , Claudine Bousquet, Nelly Lassu, Annie-Franc¸oise Maggio, Corinne Civade, Charlotte Brenier, Laurent Lempereur Agence nationale de sécurité du médicament et des produits de santé (ANSM), Direction des Contrôles (CTROL), 143/147 boulevard Anatole France, 93285 Saint-Denis Cedex, France
a r t i c l e
i n f o
Article history: Received 30 October 2014 Received in revised form 6 January 2015 Accepted 7 January 2015 Available online 19 January 2015 Keywords: Teeth bleaching products Hair products Hydrogen peroxide ISO 12787 High performance liquid chromatography/UV detection
a b s t r a c t This manuscript presents an HPLC/UV method for the determination of hydrogen peroxide present or released in teeth bleaching products and hair products. The method is based on an oxidation of triphenylphosphine into triphenylphosphine oxide by hydrogen peroxide. Triphenylphosphine oxide formed is quantified by HPLC/UV. Validation data were obtained using the ISO 12787 standard approach, particularly adapted when it is not possible to make reconstituted sample matrices. For comparative purpose, hydrogen peroxide was also determined using ceric sulfate titrimetry for both types of products. For hair products, a cross validation of both ceric titrimetric method and HPLC/UV method using the cosmetic 82/434/EEC directive (official iodometric titration method) was performed. Results obtained for 6 commercialized teeth whitening products and 5 hair products point out similar hydrogen peroxide contain using either the HPLC/UV method or ceric sulfate titrimetric method. For hair products, results were similar to the hydrogen peroxide content using the cosmetic 82/434/EEC directive method and for the HPLC/UV method, mean recoveries obtained on spiked samples, using the ISO 12787 standard, ranges from 100% to 110% with a RSD < 3.0%. To assess the analytical method proposed, the HPLC method was used to control 35 teeth bleaching products during a market survey and highlight for 5 products, hydrogen peroxide contents higher than the regulated limit. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Hydrogen peroxide (H2 O2 ) is a by-product of the metabolism in living cells. The hazards of hydrogen peroxide depend strongly on the concentration used. Hydrogen peroxide is harmful if ingested or breathed in and has low toxicity through skin contact. Depending on its concentration, hydrogen peroxide has irritant to corrosive effect on the skin, eye, and respiratory tract. Regarding tooth bleaching products, according to the Scientific Committee on Consumer Products (SCCP) opinion [1], up to 0.1% hydrogen peroxide contain the product does not pose a risk to the health of the consumer, from 0.1% to 6% hydrogen peroxide entails potential risks and because of the increasing risks of acute and long-term effects, tooth whitening products containing amount of hydrogen peroxide higher than 6% are not be considered safe for use by the consumer.
∗ Corresponding author. Tel.: +33 467064689; fax: +33 467064697. E-mail address: [email protected] (P. Gimeno). http://dx.doi.org/10.1016/j.jpba.2015.01.018 0731-7085/© 2015 Elsevier B.V. All rights reserved.
Commonly known local risks associated to hydrogen peroxide in tooth bleaching products are tooth sensitivity, gingival irritation and potential adverse effects on enamel and teeth restorative materials [2]. The council directive 2011/84/EU of 20 September 2011, which classifies teeth whitening products as cosmetics products has come into effect on 30 October 2012. This directive stipulates that only products containing less than 0.1% hydrogen peroxide, present or released, can be sold on the open market in the European Union. Products containing between 0.1% and 6% hydrogen peroxide, present or released can only be sold to dental practitioners. All teeth whitening products with more than 6% hydrogen peroxide are forbidden in the European Union. In most teeth bleaching products, the active ingredient is hydrogen peroxide or a substance able to generate hydrogen peroxide (carbamide peroxide, sodium perborate, calcium peroxide, etc.). Attention should be paid to the use of sodium perborate which is classified H360 (reprotoxic) and is regulated by the commission regulation (EC) No. 790/2009 of 10 August 2009 amending the regulation (EC) No. 1272/2008 on classification,
P. Gimeno et al. / Journal of Pharmaceutical and Biomedical Analysis 107 (2015) 386–393
labelling and packaging of substances and mixtures. For hair products, the maximum hydrogen peroxide content is limited to 12% by the regulation (EC) No. 1223/2009 of the European parliament and of the council of 30 November 2009 on cosmetic products which has replaced the cosmetics directive 76/768/EEC on 11 July of 2013. The limitation includes hydrogen peroxide present or released by other compounds or mixtures, i.e. carbamide peroxide and zinc peroxide. This regulation also limits the content of hydrogen peroxide in skin products (≤4% H2 O2 present or released) and nails hardening products (≤2% H2 O2 present or released). Its use is prohibited in other cosmetics. Up to now, the only official method for the determination of hydrogen peroxide in cosmetics is a titrimetric method described for hair-products and proposed in directive 82/434/EEC. Several other methods are available in the literature for the assay of hydrogen peroxide. These methods are based on different techniques such as: volumetric analysis, spectrometry, chromatography methods and electrochemistry. Determinations can be directly performed on hydrogen peroxide, involve redox reactions or describe intermediate preparation of coloured compounds. The most common volumetric methods for the determination of hydrogen peroxide involve either an oxidation by ceric sulfate [3,4] or by potassium permanganate (manganimetry) [5–7] or a reduction by iodide followed by sodium thiosulfate titration of the released iodine (iodometry) [5,8–11]. Another manuscript reports a protocol in which H2 O2 reduces thallium (III) into thallium (I) which is titrated bromatometrically [12]. Volumetric methods which are suitable for several substrates have been described for teeth bleaching products containing 3–11% H2 O2 present or released [3,5,8,10] and for hair products with concentrations of H2 O2 ranging from 0.01% to 9% using iodometry (official method) [9,11] or manganimetry [7]. These methods may be affected if interfering compounds are present in the cosmetic matrix. Thus the specificity towards hydrogen peroxide must be checked especially in case of products with complex matrices. Spectrometric methods for the determination of hydrogen peroxide generally involve the preparation of coloured compounds by complex formation [12,13] or by enzyme catalyzed reactions [3,14–17]. Chemiluminescence assays of H2 O2 using molecules such as luminol [18], oxalates [19] or fluorescein [20] have also been described. These methods are described for the study of both teeth bleaching products and hair-products. A method based on the reaction of hydrogen peroxide with a kit consisting of sorbitol, ferrous ammonium sulfate and xylenol has been reported for whitening strips containing 10% H2 O2 and for the measurement of residual H2 O2 after bleaching [13]. Others protocols rely on the peroxidase catalyzed reaction of hydrogen peroxide with 4-aminoantipyrine in the presence of phenolic derivatives [15,17] or with o-anisidine [16]. They have been applied to the study of teeth whitening strips containing 5% hydrogen peroxide or 10% carbamide peroxide and to the study of hair products containing between 5% and 9% H2 O2 . The drawback of such spectrometric methods, used without separation of the coloured of chemiluminescent analytes, is the lack of specificity in case of presence of compounds absorbing at the same wavelength in the solutions examined. Chromatographic techniques for the determination of hydrogen peroxide generally consist of HPLC analysis with different types of detection. In a recent manuscript of 2013, H2 O2 in aqueous solution reacts with iodide and vanillic acid to form iodovanillic acid which is quantified by HPLC/UV [21]. In another method, described for the determination of H2 O2 in disinfectants containing peracetic acid [22], H2 O2 oxidizes triphenylphosphine to yield triphenylphosphine oxide. Both compounds are analyzed using HPLC with UV. This method which requires classic HPLC equipments, is easy to set up and may be usable for the study of both teeth bleaching and hair products. The specificity of the method might however
387
be affected if other strong oxidizers susceptible of reacting with triphenylphosphine are present in the products analyzed. The use of HPLC with fluorescence detection is reported too. The techniques involve the preparation of fluorescent compounds by reaction of H2 O2 with 4-phenylacetic acid or methanol/fluoral.P in an enzyme or enzyme-like catalyzed process [23–25]. The protocols with 4phenylacetic acid require the use a post-column derivation system. HPLC with electrochemical detection allows a direct determination of hydrogen peroxide without further reaction [26]. A technique based on gas chromatography with headspace injection [27] has been reported for the determination of residual hydrogen peroxide in pulp bleaching effluents. Hydrogen peroxides react with potassium permanganate and oxygen released is analyzed with a thermal conductivity detector. For teeth bleaching products, electrochemical methods based on catalase bioassay [5], on voltammetric detection [10] or on horseradish peroxidase biosensor [14] are mentioned for contents ranging from 1% to 13% H2 O2 , present or released. For hair products, electrochemical methods, described for concentrations ranging from 0.01% to 9%, rely on modified electrodes, consisting of sensors [6,7] or biosensors [11,16], or on fluoride specific electrode [28]. Electrochemical methods are quite specific but they require special equipments as well as the preparation of the modified electrodes. In this manuscript, an HPLC with UV detection method based on an oxidation of triphenylphosphine into triphenylphosphine oxide was used for the determination of hydrogen peroxide present or released in teeth bleaching products and hair products. The proposed method could be considered as suitable for use in most laboratories (routine technique), sensitive for the quantification of low H2 O2 amount compared to titrimetric ones (0.1%) most often used for the assay of cosmetics and specific enough provided the list of ingredients has been checked. Indeed, less strong oxidizers compounds are present in teeth whitening products or hair cosmetics compared to disinfectants. Validation data were obtained using the ISO 12787 standard approach for cosmetics [29]. For both types of products, the ceric sulfate titrimetric, method described for a wide range of matrices including hair-products and teeth bleaching products [3,4], was used as a comparative method. Indeed, titrimetric methods can be considered as reference methods after a previous checking of their specificity towards the different matrices selected. For hair products, a cross validation of both analytical method (ceric sulfate and HPLC/UV) using the only official method for the assay H2 O2 in cosmetics (82/434/EEC directive based on iodometric titration [9]) was performed. To our knowledge, it is the first time that an HPLC/UV method is reported for the analysis of H2 O2 in complex matrices like cosmetics. Results obtained for 35 teeth bleaching products controlled during a market survey are given to assess the analytical method proposed.
2. Materials and methods 2.1. Reagents Standards with purity equal to or greater than, triphenylphosphine 99%, hydrogen peroxide 30% solution in water, urea peroxide 97%, sodium thiosulfate 98.0%, potassium iodide 99.5%, 1,10phenanthroline monohydrate 99%, sulphuric acid 95–98% and ferrous sulfate 99% were purchased from Sigma–Aldrich (St. Quentin Fallavier, France). 0.1 N ceric (IV) sulfate and starch (soluble grade for analysis) comes from Merk (Fontenay sous bois, France). Ammonium molybdate (RPE-ACS-Reagent) comes from Carlo Erba (Val de Reuil, France) and HPLC grade acetonitrile was obtained from J. T. Baker (Deventer, Netherlands). Ultra-pure grade water (Milli-Q, Millipore) was used for all assays.
388
P. Gimeno et al. / Journal of Pharmaceutical and Biomedical Analysis 107 (2015) 386–393
Table 1 Solvent gradient used for the analysis of hydrogen peroxide by HPLC/UV. Time
H2 O (%) v/v
CH3 CN (%) v/v
0 min 5.5 min 6.5 min 9.0 min 10.0 min 20.0 min
50 50 0 0 50 50
50 50 100 100 50 50
2.2. Cosmetic samples 2.2.1. Teeth whitening products For the determination of some validation data on the method used according to the ISO 12787, 6 teeth whitening products were analyzed: 5 products containing hydrogen peroxide (3 paint on gels – gel A, gel B, gel C; 1 gel coated on a plastic strip – strips 1; 1 gel impregnated on a foam thick strip – strips 2) and 1 product containing carbamide peroxide (gel D). The absence of potentially interfering compound for instance oxidative substances, such as chlorite or metabisulphite salt, which might be present in these cosmetics was checked in the list of ingredient. To assess the analytical method proposed, the HPLC method was used to control 35 teeth bleaching products during a market survey, among which 20 sold on the open market and 15 sold to dental practitioners. These kits contained a least one whitening formulation (gel or foam) and in some cases other product parts. Some of these products were analyzed (activators, toothpastes, mouthwash, whitening cream, repairing serum, etc.), others were not analyzed (cheek retractor, cotton rolls, barrier material, face bib, masking cream, gauze, etc.). 2.2.2. Hair products 5 hair products: 4 oxidative hair dyes (HD-1 to 4) and 1 discoloration spray (Sp-5) were assayed using the ISO 12787. As for teeth whitening products, the absence of interfering compounds like chlorite salt (sodium chlorite) was checked before testing. 2.3. HPLC/UV method All analyses were performed on an Agilent 1200 HPLC system (Santa Clara, CA, USA) equipped with a diode array detector. The column was a reverse phase Nucleosil® C18 Macherey-Nagel® (Düren, ˚ The injection volume Germany) 150 mm × 4.6 mm, 5 m, 100 A.
was 10 l and detection was performed at 225 nm. The mobile phase consisted of a gradient elution of water and acetonitrile (see Table 1). The column temperature was set at 25 ◦ C and the mobile phase flow rate was set at 1.0 mL min−1 . An example of chromatogram corresponding to 120 g mL−1 H2 O2 is shown in Fig. 1. 2.4. Standard solutions preparation Hydrogen peroxide 30% or urea peroxide standards can be used for the preparation of H2 O2 standard solutions. A 1000 g mL−1 standard solution of H2 O2 is first prepared either diluting 0.167 g of hydrogen peroxide 30% solution in 50 mL of water or weighing 140 mg of urea peroxide (equivalent to 36% H2 O2 ) in 50 mL volumetric flask with water. From the previous standard solution (1000 g mL−1 ), at least 5 standard solutions ranging from 30 to 200 g mL−1 H2 O2 are prepared by dilutions in water/acetonitrile (35/65, v/v). Depending on sample preparations, this calibration range allows to assay percent of H2 O2 in real products ranging from 0.075% to 10% for teeth whitening products and from 0.03% to 16% for hair products. These values are in agreement with maximal contents of hydrogen peroxide regulated in these types of products. In order to check the system apparatus conformity during sample run analysis, an independent quality control standard solution at 100 g mL−1 is steadily injected (biases < 5%, in agreement with in-house specification). 2.5. Samples preparation 0.1–2.0 g of gel are weighted and dissolved in 50 mL of water. For impregnated strips, the exact mass of gel present on the strip must be first determined weighting a strip before and after a complete extraction of the gel in water. For hair products, 0.5–1.0 g of sample is weighted in a centrifugation tube and dissolved in 10–400 mL of water. The preparation is shake for 1 min and centrifuged. For oxidative hair dyes (HD-1 to 4), the sample is first prepared by mixing the cosmetic cream and the oxidative solution according to instructions for use. The hydrogen peroxide content in the sample is assayed using the HPLC/UV method. Sample weights and water volumes are adapted to the H2 O2 amount present, in order to have a final H2 O2 injected vial concentration included into the linearity range (30–200 g mL−1 ). For titrimetric methods, samples are weighted directly in the flask and assay performed according to well-known protocols using ceric sulphate [3,4] or iodometry [9].
Fig. 1. Chromatogram obtained for a 120 g mL−1 H2 O2 standard solution.
P. Gimeno et al. / Journal of Pharmaceutical and Biomedical Analysis 107 (2015) 386–393
Fig. 2. Formula used for the determination of H2 O2 by titrimetric methods in the cosmetic sample.
The amount of H2 O2 present in the cosmetic sample was directly calculated using the formula given in Fig. 2. 2.6. Reaction of H2 O2 with triphenylphosphine Under common conditions triphenylphosphine (TPP) is slightly oxidized by molecular oxygen. The formation of triphenylphosphine oxide (TPPO) from TPP needs the use of strong oxidants. In this manuscript, TPPO is obtained from TPP by an oxidative reaction using hydrogen peroxide (H2 O2 ) as a moderately strong oxidant [30]. A solution of TPP is prepared at 0.01 M in acetonitrile weighing 65.5 mg of TPP in a 25 mL volumetric brown flask. In a 10 mL centrifugal tube protected from light 5 mL of acetonitrile, 3 mL of ultra pure water and 1 mL of the TPP solution previously prepared are introduced. The mixture is stirred using a rotary agitator. 1 mL of the solution to be examined (hydrogen peroxide standard solution (2.4) or of sample solution (2.5)) is added to the tube and the solution is stirred again. The sample preparation is then left standing at least 2 h before been injected into HPLC. For a complete reaction of H2 O2 present, TPP must be present in excess during reaction. The reaction of all H2 O2 present is highlight by the presence of a residual peak of TPP (un-reacted) in the HPLC chromatograms. Attention should be paid to the presence of small amount of TPPO in TPP standard coming from an unavoidable residual oxidation. Blank reaction must be performed, in order to estimate the contribution of this residual TPPO on the final result, using 1 mL of ultra pure water instead of the solution to be examined. Otherwise, its presence may lead to a systematic error in the measurement of TPPO coming from the oxidation of TPP by H2 O2 present in the sample. Determination of H2 O2 is made by external quantification using TPPO peaks area. 2.7. Validation strategy 2.7.1. ISO 12787 international standard The ISO 12787 international standard defines validation criteria to which analytical results obtained from the analysis of cosmetic products should comply in order to give confidence in performance, reliability and quality of the final result. It proposes an analytical approach for chromatographic analyses in order to obtain at the same time an assay result and different validation elements which can be determined for each sample (or sample groups) submitted to the analysis. This approach allows the obtention of validation criteria on the assay results taking into account the cosmetic matrix. This standard is particularly recommended for the determination of some validation data on the assay when the method have to be applied to different cosmetic matrices in the absence of blank formulations or reference samples. The ISO standard is divided into three parts. The first part consists in determining the main characteristics of the analytical method used on standard solutions before performing tests on samples. Values obtained could be considered as best ones achievable in cosmetics. Validation criteria checked during this first step can be if relevant: limit of detection (LoD)
389
and/or limit of quantification (LoQ); conformity of the chromatographic analysis (Rs: Resolution, As: Asymmetry); linear range of the analyte signal and standard accuracy (biases). The second and third parts of this standard, called “screening step” and “assay” focused on matrices, consists respectively in evaluating the quantity of analyte in the sample, if present, in order to adapt the sample assay protocol and determining the analyte amount using spiked preparations. These parts are used in order to check the matrix specificity and matrix effects on the sample assayed when no blank or reference samples are available (i.e. LoQ in matrix, compound interferences, recoveries, etc.). 2.7.2. Validation protocol As most teeth whitening products have similar compositions (water, alcohols, polyols, gel forming agents, viscosity controllers, buffering agents or emulsion stabilizing compounds) and aspects, after a previous verification of the absence of a possible matrix effect, only the first part of this ISO 12787 standard was performed. Using standard solutions (see Section 3.2.1), different validation parameters like the method suitability, the stability of standard solution, the intra- and inter-day repeatability and the quantification limits were determined using the ISO approach [29]. The linearity between the TPPO concentration and the area of the corresponding peak was checked for TPPO solutions obtained by reaction between TPP and H2 O2 (TPP + H2 O2 ) and for TPPO solutions obtained by reaction between TPP and urea peroxide (TPP + urea peroxide). A comparison of the curves obtained from TPP + H2 O2 and from TPP + urea peroxide was performed in order to check if the method is suitable for both types of products containing H2 O2 or urea peroxide. For hair cosmetic products, as composition may change depending on the type of product analyzed (shampoo, hair dyes, lotions, etc.) and in the absence of blank/reference samples representative of the different matrices formulations, all the ISO 12787 standard was considered, including the 2nd and 3rd steps using spiked samples (see Section 3.2.2). For all analyzed samples, a common H2 O2 spiked amount fixed at 1% was used due to the presence in these cosmetics of hydrogen peroxide amount ranging from 2% to 4% (evaluated during the 2nd step of the ISO 12787). For teeth bleaching and hair products, the specificity and the accuracy of the method were checked by comparing results obtained using the HPLC/UV method to results obtained on the same products using ceric sulfate titration (see Section 3.2.3). For hair products, an additional cross validation of both methods (HPLC–UV and ceric sulfate titration) using the cosmetic 82/434/EEC directive as reference method was also performed (see Section 3.2.3.2). Moreover, for both products the absence of TPPO co-eluted compounds (interference in the sample) was checked by the injection of a non-derivative sample preparation (sample preparation injected without the TPP reaction step – see Section 2.6). Statistic processing of the data was performed using the AVA (Aide à la Validation Analytique) software version 3.1 distributed by QUALILAB. 3. Results and discussions 3.1. Preliminary studies During the method development, 4 methods were initially selected and tested for the quantification of hydrogen peroxide in teeth whitening products: 3 titrimetric methods using respectively potassium permanganate, sodium thiosulfate and ceric sulfate and a chromatographic method implementing a reaction between hydrogen peroxide and the triphenyl phosphine (TPP) to form the corresponding oxide (TPPO). During preliminary assays, a lack of specificity was pointed out for all teeth whitening products with the manganimetric method. Indeed, alcohols such as glycerol,
390
P. Gimeno et al. / Journal of Pharmaceutical and Biomedical Analysis 107 (2015) 386–393
Table 2 Validation data on standard solutions obtained for HPLC–UV method using the ISO 12787. Validation criteria on standard solution (according to ISO 12787) Stability
Standard solutions (H2 O2 and TPP)
>24 h
RSD injection precision TPPO Symmetry (As) Resolution (TPP/TPPO) (Rs)
0.4% 1.04 >1.5
Quantification Limit (LoQ) (based on TPPO peak area)
Value
30 g mL−1 H2 O2
Linearity (H2 O2 + TPP)
Linearity range Determination coefficient R2 Variance Homogeneity (COCHRAN test) Significant slope existence (FISHER) Validity of calibration curve (FISHER) Y-intercept comparison with 0 (STUDENT) Biases
[30–200 g mL−1 ] H2 O2 >0.9995 NS* HS** NS* S*** 1.5 according to the Ph. Eur. 2.2.46) and a number of theoretical plates (N) higher than 10,000 for the TPPO peak. Furthermore, no carryover was pointed out during assay. According to
these results, the chromatographic method can be considered as suitable for the analysis of H2 O2 . Due to the presence of a residual peak of TPPO in TPP standard, a TPPO peak is often present in blank chromatograms involving the necessity to carry out blank preparations for all assays. 3.2.1.2. Stability of standard solutions. The stability of TPP + H2 O2 solutions was checked by re-injecting the six solutions prepared on the first day of the precision study into the HPLC system after 24 h of storage in the HPLC autosampler (dark + 6 ◦ C). The RSD and the confidence interval were determined for both series of values (first injected solutions and re-injected ones). Moreover, the area evolution in 24 h was calculated for each type of solution studied. Results obtained highlight an area increase of the solutions studied in 24 h comprised between 1.5% and 1.8% with an average value of 1.7%. The confidence interval calculated on the areas ranges from 5404 to 5513 with a RSD of 1.2% at T0 and from 5440 to 5603 with a RSD of 1.2% at T0 + 24 h. The overlap of both confidence intervals, respectively calculated at T0 and at T0 + 24 h, and the RSD calculated at 1.5% by grouping the areas obtained after preparation and the areas obtained after 24 h clearly highlight that the area variations observed in 24 h could be considered as negligible. Solutions of TPPO obtained after reaction between H2 O2 and TPP were considered as stable over a period of 24 h if stored in the dark at + 6 ◦ C (i.e. HPLC autosampler). The stability of TPPO was checked and confirmed on prepared cosmetic samples. 3.2.1.3. Quantification limit. The quantification limit (LoQ), corresponding to the lower H2 O2 concentration in a standard solution that could be quantified accurately, was set to 30 g mL−1 corresponding to 0.03–0.07% H2 O2 in cosmetics depending on sample preparation. This limit takes into account the influence of residual TPPO present in TPP standard and allows quantification of H2 O2 at the target value set to 0.1% by the cosmetic regulation in teeth whitening products. This limit is the best one achievable in cosmetics. Depending matrices, this limit could be higher and, if relevant, must be checked according to the ISO 12787 standard (steps 2 and 3). 3.2.1.4. Linearity. Calibration curves were performed three nonconsecutive days using fresh standard solution. Results of the linearity study for the comparison between the curves of
P. Gimeno et al. / Journal of Pharmaceutical and Biomedical Analysis 107 (2015) 386–393 Table 3 Results obtained for the linearity study of hydrogen peroxide and urea peroxide. H2 O2 + TPP
Urea H2 O2 + TPP
Number of calibration level Calibration range Variance homogeneity (COCHRAN test) Aberrant values (DIXON test)
8 [30–200 mg L–1 ] NS*
8 [30–200 mg L–1 ] NS*
No
No
Slope Y-intercept Determination coefficient (R2 ) Concentration deviation of the first standard calibration level (%)
49.44 388.50 0.99973 Less or near 5%
49.72 293.02 0.99957 Less or near 5%
Slope existence (FISHER test) Validity of the calibration curve (FISHER test) Y-intercept comparison with zero (STUDENT test)
HS** NS*
HS** NS*
S***
S***
Slopes comparison Y-intercept comparison * ** ***
NS* NS*
No significant difference (NS). Highly significant difference (HS). Significant difference (S).
TPP + H2 O2 and the curves TPP + urea peroxide using the AVA software are presented in Table 3. The variance homogeneity calculated on calibration values is non-significant (homoscedastic). This allows a valid statistical process of all the linearity criteria. For all curves the determination coefficient (R2 ) is higher than 0.9995 and the bias obtained for the 1st concentration level is lower than 5%. These results display a good linearity correlation for all solutions on the calibration range considered (30–200 g mL−1 ). Y-intercept comparison with zero (STUDENT) is significant in all cases. This might be explained by the fact that commercially available TPP contains approximately 1% TPPO. Comparison between the slopes and Y-intercepts of the curves respectively obtained from hydrogen peroxide solutions and from urea peroxide solutions are non-significant. The two curves can be considered as similar. The results obtained for the determination of hydrogen peroxide and of urea peroxide are comparable. Both, hydrogen peroxide or urea peroxide can be used for standard calibration. 3.2.1.5. Intra- and inter-day precision/residues. Six independent solutions containing each 100 g mL−1 of hydrogen peroxide were prepared everyday during three different days from the corresponding stock solution H2 O2 (prepared every day from commercial H2 O2 ). The intra- and inter-days RSD calculated are both around 1.3% and can thus be considered as acceptable. The nonsignificant variance homogeneity allows a valid statistical process of the intra- and inter-days precision study. The confidence interval calculated on the mean recovery, ranging form 99.5% to 100.9%, is less than 1.5% wide and includes the 100% value. The accuracy of the method can be considered as acceptable. Moreover the fact that the mean validity test is non-significant indicates that there is no day effect. This result meets for example the accuracy requirement for pharmaceutical assay (
Contents lists available at ScienceDirect
Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba
High-performance liquid chromatography method for the determination of hydrogen peroxide present or released in teeth bleaching kits and hair cosmetic products Pascal Gimeno ∗ , Claudine Bousquet, Nelly Lassu, Annie-Franc¸oise Maggio, Corinne Civade, Charlotte Brenier, Laurent Lempereur Agence nationale de sécurité du médicament et des produits de santé (ANSM), Direction des Contrôles (CTROL), 143/147 boulevard Anatole France, 93285 Saint-Denis Cedex, France
a r t i c l e
i n f o
Article history: Received 30 October 2014 Received in revised form 6 January 2015 Accepted 7 January 2015 Available online 19 January 2015 Keywords: Teeth bleaching products Hair products Hydrogen peroxide ISO 12787 High performance liquid chromatography/UV detection
a b s t r a c t This manuscript presents an HPLC/UV method for the determination of hydrogen peroxide present or released in teeth bleaching products and hair products. The method is based on an oxidation of triphenylphosphine into triphenylphosphine oxide by hydrogen peroxide. Triphenylphosphine oxide formed is quantified by HPLC/UV. Validation data were obtained using the ISO 12787 standard approach, particularly adapted when it is not possible to make reconstituted sample matrices. For comparative purpose, hydrogen peroxide was also determined using ceric sulfate titrimetry for both types of products. For hair products, a cross validation of both ceric titrimetric method and HPLC/UV method using the cosmetic 82/434/EEC directive (official iodometric titration method) was performed. Results obtained for 6 commercialized teeth whitening products and 5 hair products point out similar hydrogen peroxide contain using either the HPLC/UV method or ceric sulfate titrimetric method. For hair products, results were similar to the hydrogen peroxide content using the cosmetic 82/434/EEC directive method and for the HPLC/UV method, mean recoveries obtained on spiked samples, using the ISO 12787 standard, ranges from 100% to 110% with a RSD < 3.0%. To assess the analytical method proposed, the HPLC method was used to control 35 teeth bleaching products during a market survey and highlight for 5 products, hydrogen peroxide contents higher than the regulated limit. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Hydrogen peroxide (H2 O2 ) is a by-product of the metabolism in living cells. The hazards of hydrogen peroxide depend strongly on the concentration used. Hydrogen peroxide is harmful if ingested or breathed in and has low toxicity through skin contact. Depending on its concentration, hydrogen peroxide has irritant to corrosive effect on the skin, eye, and respiratory tract. Regarding tooth bleaching products, according to the Scientific Committee on Consumer Products (SCCP) opinion [1], up to 0.1% hydrogen peroxide contain the product does not pose a risk to the health of the consumer, from 0.1% to 6% hydrogen peroxide entails potential risks and because of the increasing risks of acute and long-term effects, tooth whitening products containing amount of hydrogen peroxide higher than 6% are not be considered safe for use by the consumer.
∗ Corresponding author. Tel.: +33 467064689; fax: +33 467064697. E-mail address: [email protected] (P. Gimeno). http://dx.doi.org/10.1016/j.jpba.2015.01.018 0731-7085/© 2015 Elsevier B.V. All rights reserved.
Commonly known local risks associated to hydrogen peroxide in tooth bleaching products are tooth sensitivity, gingival irritation and potential adverse effects on enamel and teeth restorative materials [2]. The council directive 2011/84/EU of 20 September 2011, which classifies teeth whitening products as cosmetics products has come into effect on 30 October 2012. This directive stipulates that only products containing less than 0.1% hydrogen peroxide, present or released, can be sold on the open market in the European Union. Products containing between 0.1% and 6% hydrogen peroxide, present or released can only be sold to dental practitioners. All teeth whitening products with more than 6% hydrogen peroxide are forbidden in the European Union. In most teeth bleaching products, the active ingredient is hydrogen peroxide or a substance able to generate hydrogen peroxide (carbamide peroxide, sodium perborate, calcium peroxide, etc.). Attention should be paid to the use of sodium perborate which is classified H360 (reprotoxic) and is regulated by the commission regulation (EC) No. 790/2009 of 10 August 2009 amending the regulation (EC) No. 1272/2008 on classification,
P. Gimeno et al. / Journal of Pharmaceutical and Biomedical Analysis 107 (2015) 386–393
labelling and packaging of substances and mixtures. For hair products, the maximum hydrogen peroxide content is limited to 12% by the regulation (EC) No. 1223/2009 of the European parliament and of the council of 30 November 2009 on cosmetic products which has replaced the cosmetics directive 76/768/EEC on 11 July of 2013. The limitation includes hydrogen peroxide present or released by other compounds or mixtures, i.e. carbamide peroxide and zinc peroxide. This regulation also limits the content of hydrogen peroxide in skin products (≤4% H2 O2 present or released) and nails hardening products (≤2% H2 O2 present or released). Its use is prohibited in other cosmetics. Up to now, the only official method for the determination of hydrogen peroxide in cosmetics is a titrimetric method described for hair-products and proposed in directive 82/434/EEC. Several other methods are available in the literature for the assay of hydrogen peroxide. These methods are based on different techniques such as: volumetric analysis, spectrometry, chromatography methods and electrochemistry. Determinations can be directly performed on hydrogen peroxide, involve redox reactions or describe intermediate preparation of coloured compounds. The most common volumetric methods for the determination of hydrogen peroxide involve either an oxidation by ceric sulfate [3,4] or by potassium permanganate (manganimetry) [5–7] or a reduction by iodide followed by sodium thiosulfate titration of the released iodine (iodometry) [5,8–11]. Another manuscript reports a protocol in which H2 O2 reduces thallium (III) into thallium (I) which is titrated bromatometrically [12]. Volumetric methods which are suitable for several substrates have been described for teeth bleaching products containing 3–11% H2 O2 present or released [3,5,8,10] and for hair products with concentrations of H2 O2 ranging from 0.01% to 9% using iodometry (official method) [9,11] or manganimetry [7]. These methods may be affected if interfering compounds are present in the cosmetic matrix. Thus the specificity towards hydrogen peroxide must be checked especially in case of products with complex matrices. Spectrometric methods for the determination of hydrogen peroxide generally involve the preparation of coloured compounds by complex formation [12,13] or by enzyme catalyzed reactions [3,14–17]. Chemiluminescence assays of H2 O2 using molecules such as luminol [18], oxalates [19] or fluorescein [20] have also been described. These methods are described for the study of both teeth bleaching products and hair-products. A method based on the reaction of hydrogen peroxide with a kit consisting of sorbitol, ferrous ammonium sulfate and xylenol has been reported for whitening strips containing 10% H2 O2 and for the measurement of residual H2 O2 after bleaching [13]. Others protocols rely on the peroxidase catalyzed reaction of hydrogen peroxide with 4-aminoantipyrine in the presence of phenolic derivatives [15,17] or with o-anisidine [16]. They have been applied to the study of teeth whitening strips containing 5% hydrogen peroxide or 10% carbamide peroxide and to the study of hair products containing between 5% and 9% H2 O2 . The drawback of such spectrometric methods, used without separation of the coloured of chemiluminescent analytes, is the lack of specificity in case of presence of compounds absorbing at the same wavelength in the solutions examined. Chromatographic techniques for the determination of hydrogen peroxide generally consist of HPLC analysis with different types of detection. In a recent manuscript of 2013, H2 O2 in aqueous solution reacts with iodide and vanillic acid to form iodovanillic acid which is quantified by HPLC/UV [21]. In another method, described for the determination of H2 O2 in disinfectants containing peracetic acid [22], H2 O2 oxidizes triphenylphosphine to yield triphenylphosphine oxide. Both compounds are analyzed using HPLC with UV. This method which requires classic HPLC equipments, is easy to set up and may be usable for the study of both teeth bleaching and hair products. The specificity of the method might however
387
be affected if other strong oxidizers susceptible of reacting with triphenylphosphine are present in the products analyzed. The use of HPLC with fluorescence detection is reported too. The techniques involve the preparation of fluorescent compounds by reaction of H2 O2 with 4-phenylacetic acid or methanol/fluoral.P in an enzyme or enzyme-like catalyzed process [23–25]. The protocols with 4phenylacetic acid require the use a post-column derivation system. HPLC with electrochemical detection allows a direct determination of hydrogen peroxide without further reaction [26]. A technique based on gas chromatography with headspace injection [27] has been reported for the determination of residual hydrogen peroxide in pulp bleaching effluents. Hydrogen peroxides react with potassium permanganate and oxygen released is analyzed with a thermal conductivity detector. For teeth bleaching products, electrochemical methods based on catalase bioassay [5], on voltammetric detection [10] or on horseradish peroxidase biosensor [14] are mentioned for contents ranging from 1% to 13% H2 O2 , present or released. For hair products, electrochemical methods, described for concentrations ranging from 0.01% to 9%, rely on modified electrodes, consisting of sensors [6,7] or biosensors [11,16], or on fluoride specific electrode [28]. Electrochemical methods are quite specific but they require special equipments as well as the preparation of the modified electrodes. In this manuscript, an HPLC with UV detection method based on an oxidation of triphenylphosphine into triphenylphosphine oxide was used for the determination of hydrogen peroxide present or released in teeth bleaching products and hair products. The proposed method could be considered as suitable for use in most laboratories (routine technique), sensitive for the quantification of low H2 O2 amount compared to titrimetric ones (0.1%) most often used for the assay of cosmetics and specific enough provided the list of ingredients has been checked. Indeed, less strong oxidizers compounds are present in teeth whitening products or hair cosmetics compared to disinfectants. Validation data were obtained using the ISO 12787 standard approach for cosmetics [29]. For both types of products, the ceric sulfate titrimetric, method described for a wide range of matrices including hair-products and teeth bleaching products [3,4], was used as a comparative method. Indeed, titrimetric methods can be considered as reference methods after a previous checking of their specificity towards the different matrices selected. For hair products, a cross validation of both analytical method (ceric sulfate and HPLC/UV) using the only official method for the assay H2 O2 in cosmetics (82/434/EEC directive based on iodometric titration [9]) was performed. To our knowledge, it is the first time that an HPLC/UV method is reported for the analysis of H2 O2 in complex matrices like cosmetics. Results obtained for 35 teeth bleaching products controlled during a market survey are given to assess the analytical method proposed.
2. Materials and methods 2.1. Reagents Standards with purity equal to or greater than, triphenylphosphine 99%, hydrogen peroxide 30% solution in water, urea peroxide 97%, sodium thiosulfate 98.0%, potassium iodide 99.5%, 1,10phenanthroline monohydrate 99%, sulphuric acid 95–98% and ferrous sulfate 99% were purchased from Sigma–Aldrich (St. Quentin Fallavier, France). 0.1 N ceric (IV) sulfate and starch (soluble grade for analysis) comes from Merk (Fontenay sous bois, France). Ammonium molybdate (RPE-ACS-Reagent) comes from Carlo Erba (Val de Reuil, France) and HPLC grade acetonitrile was obtained from J. T. Baker (Deventer, Netherlands). Ultra-pure grade water (Milli-Q, Millipore) was used for all assays.
388
P. Gimeno et al. / Journal of Pharmaceutical and Biomedical Analysis 107 (2015) 386–393
Table 1 Solvent gradient used for the analysis of hydrogen peroxide by HPLC/UV. Time
H2 O (%) v/v
CH3 CN (%) v/v
0 min 5.5 min 6.5 min 9.0 min 10.0 min 20.0 min
50 50 0 0 50 50
50 50 100 100 50 50
2.2. Cosmetic samples 2.2.1. Teeth whitening products For the determination of some validation data on the method used according to the ISO 12787, 6 teeth whitening products were analyzed: 5 products containing hydrogen peroxide (3 paint on gels – gel A, gel B, gel C; 1 gel coated on a plastic strip – strips 1; 1 gel impregnated on a foam thick strip – strips 2) and 1 product containing carbamide peroxide (gel D). The absence of potentially interfering compound for instance oxidative substances, such as chlorite or metabisulphite salt, which might be present in these cosmetics was checked in the list of ingredient. To assess the analytical method proposed, the HPLC method was used to control 35 teeth bleaching products during a market survey, among which 20 sold on the open market and 15 sold to dental practitioners. These kits contained a least one whitening formulation (gel or foam) and in some cases other product parts. Some of these products were analyzed (activators, toothpastes, mouthwash, whitening cream, repairing serum, etc.), others were not analyzed (cheek retractor, cotton rolls, barrier material, face bib, masking cream, gauze, etc.). 2.2.2. Hair products 5 hair products: 4 oxidative hair dyes (HD-1 to 4) and 1 discoloration spray (Sp-5) were assayed using the ISO 12787. As for teeth whitening products, the absence of interfering compounds like chlorite salt (sodium chlorite) was checked before testing. 2.3. HPLC/UV method All analyses were performed on an Agilent 1200 HPLC system (Santa Clara, CA, USA) equipped with a diode array detector. The column was a reverse phase Nucleosil® C18 Macherey-Nagel® (Düren, ˚ The injection volume Germany) 150 mm × 4.6 mm, 5 m, 100 A.
was 10 l and detection was performed at 225 nm. The mobile phase consisted of a gradient elution of water and acetonitrile (see Table 1). The column temperature was set at 25 ◦ C and the mobile phase flow rate was set at 1.0 mL min−1 . An example of chromatogram corresponding to 120 g mL−1 H2 O2 is shown in Fig. 1. 2.4. Standard solutions preparation Hydrogen peroxide 30% or urea peroxide standards can be used for the preparation of H2 O2 standard solutions. A 1000 g mL−1 standard solution of H2 O2 is first prepared either diluting 0.167 g of hydrogen peroxide 30% solution in 50 mL of water or weighing 140 mg of urea peroxide (equivalent to 36% H2 O2 ) in 50 mL volumetric flask with water. From the previous standard solution (1000 g mL−1 ), at least 5 standard solutions ranging from 30 to 200 g mL−1 H2 O2 are prepared by dilutions in water/acetonitrile (35/65, v/v). Depending on sample preparations, this calibration range allows to assay percent of H2 O2 in real products ranging from 0.075% to 10% for teeth whitening products and from 0.03% to 16% for hair products. These values are in agreement with maximal contents of hydrogen peroxide regulated in these types of products. In order to check the system apparatus conformity during sample run analysis, an independent quality control standard solution at 100 g mL−1 is steadily injected (biases < 5%, in agreement with in-house specification). 2.5. Samples preparation 0.1–2.0 g of gel are weighted and dissolved in 50 mL of water. For impregnated strips, the exact mass of gel present on the strip must be first determined weighting a strip before and after a complete extraction of the gel in water. For hair products, 0.5–1.0 g of sample is weighted in a centrifugation tube and dissolved in 10–400 mL of water. The preparation is shake for 1 min and centrifuged. For oxidative hair dyes (HD-1 to 4), the sample is first prepared by mixing the cosmetic cream and the oxidative solution according to instructions for use. The hydrogen peroxide content in the sample is assayed using the HPLC/UV method. Sample weights and water volumes are adapted to the H2 O2 amount present, in order to have a final H2 O2 injected vial concentration included into the linearity range (30–200 g mL−1 ). For titrimetric methods, samples are weighted directly in the flask and assay performed according to well-known protocols using ceric sulphate [3,4] or iodometry [9].
Fig. 1. Chromatogram obtained for a 120 g mL−1 H2 O2 standard solution.
P. Gimeno et al. / Journal of Pharmaceutical and Biomedical Analysis 107 (2015) 386–393
Fig. 2. Formula used for the determination of H2 O2 by titrimetric methods in the cosmetic sample.
The amount of H2 O2 present in the cosmetic sample was directly calculated using the formula given in Fig. 2. 2.6. Reaction of H2 O2 with triphenylphosphine Under common conditions triphenylphosphine (TPP) is slightly oxidized by molecular oxygen. The formation of triphenylphosphine oxide (TPPO) from TPP needs the use of strong oxidants. In this manuscript, TPPO is obtained from TPP by an oxidative reaction using hydrogen peroxide (H2 O2 ) as a moderately strong oxidant [30]. A solution of TPP is prepared at 0.01 M in acetonitrile weighing 65.5 mg of TPP in a 25 mL volumetric brown flask. In a 10 mL centrifugal tube protected from light 5 mL of acetonitrile, 3 mL of ultra pure water and 1 mL of the TPP solution previously prepared are introduced. The mixture is stirred using a rotary agitator. 1 mL of the solution to be examined (hydrogen peroxide standard solution (2.4) or of sample solution (2.5)) is added to the tube and the solution is stirred again. The sample preparation is then left standing at least 2 h before been injected into HPLC. For a complete reaction of H2 O2 present, TPP must be present in excess during reaction. The reaction of all H2 O2 present is highlight by the presence of a residual peak of TPP (un-reacted) in the HPLC chromatograms. Attention should be paid to the presence of small amount of TPPO in TPP standard coming from an unavoidable residual oxidation. Blank reaction must be performed, in order to estimate the contribution of this residual TPPO on the final result, using 1 mL of ultra pure water instead of the solution to be examined. Otherwise, its presence may lead to a systematic error in the measurement of TPPO coming from the oxidation of TPP by H2 O2 present in the sample. Determination of H2 O2 is made by external quantification using TPPO peaks area. 2.7. Validation strategy 2.7.1. ISO 12787 international standard The ISO 12787 international standard defines validation criteria to which analytical results obtained from the analysis of cosmetic products should comply in order to give confidence in performance, reliability and quality of the final result. It proposes an analytical approach for chromatographic analyses in order to obtain at the same time an assay result and different validation elements which can be determined for each sample (or sample groups) submitted to the analysis. This approach allows the obtention of validation criteria on the assay results taking into account the cosmetic matrix. This standard is particularly recommended for the determination of some validation data on the assay when the method have to be applied to different cosmetic matrices in the absence of blank formulations or reference samples. The ISO standard is divided into three parts. The first part consists in determining the main characteristics of the analytical method used on standard solutions before performing tests on samples. Values obtained could be considered as best ones achievable in cosmetics. Validation criteria checked during this first step can be if relevant: limit of detection (LoD)
389
and/or limit of quantification (LoQ); conformity of the chromatographic analysis (Rs: Resolution, As: Asymmetry); linear range of the analyte signal and standard accuracy (biases). The second and third parts of this standard, called “screening step” and “assay” focused on matrices, consists respectively in evaluating the quantity of analyte in the sample, if present, in order to adapt the sample assay protocol and determining the analyte amount using spiked preparations. These parts are used in order to check the matrix specificity and matrix effects on the sample assayed when no blank or reference samples are available (i.e. LoQ in matrix, compound interferences, recoveries, etc.). 2.7.2. Validation protocol As most teeth whitening products have similar compositions (water, alcohols, polyols, gel forming agents, viscosity controllers, buffering agents or emulsion stabilizing compounds) and aspects, after a previous verification of the absence of a possible matrix effect, only the first part of this ISO 12787 standard was performed. Using standard solutions (see Section 3.2.1), different validation parameters like the method suitability, the stability of standard solution, the intra- and inter-day repeatability and the quantification limits were determined using the ISO approach [29]. The linearity between the TPPO concentration and the area of the corresponding peak was checked for TPPO solutions obtained by reaction between TPP and H2 O2 (TPP + H2 O2 ) and for TPPO solutions obtained by reaction between TPP and urea peroxide (TPP + urea peroxide). A comparison of the curves obtained from TPP + H2 O2 and from TPP + urea peroxide was performed in order to check if the method is suitable for both types of products containing H2 O2 or urea peroxide. For hair cosmetic products, as composition may change depending on the type of product analyzed (shampoo, hair dyes, lotions, etc.) and in the absence of blank/reference samples representative of the different matrices formulations, all the ISO 12787 standard was considered, including the 2nd and 3rd steps using spiked samples (see Section 3.2.2). For all analyzed samples, a common H2 O2 spiked amount fixed at 1% was used due to the presence in these cosmetics of hydrogen peroxide amount ranging from 2% to 4% (evaluated during the 2nd step of the ISO 12787). For teeth bleaching and hair products, the specificity and the accuracy of the method were checked by comparing results obtained using the HPLC/UV method to results obtained on the same products using ceric sulfate titration (see Section 3.2.3). For hair products, an additional cross validation of both methods (HPLC–UV and ceric sulfate titration) using the cosmetic 82/434/EEC directive as reference method was also performed (see Section 3.2.3.2). Moreover, for both products the absence of TPPO co-eluted compounds (interference in the sample) was checked by the injection of a non-derivative sample preparation (sample preparation injected without the TPP reaction step – see Section 2.6). Statistic processing of the data was performed using the AVA (Aide à la Validation Analytique) software version 3.1 distributed by QUALILAB. 3. Results and discussions 3.1. Preliminary studies During the method development, 4 methods were initially selected and tested for the quantification of hydrogen peroxide in teeth whitening products: 3 titrimetric methods using respectively potassium permanganate, sodium thiosulfate and ceric sulfate and a chromatographic method implementing a reaction between hydrogen peroxide and the triphenyl phosphine (TPP) to form the corresponding oxide (TPPO). During preliminary assays, a lack of specificity was pointed out for all teeth whitening products with the manganimetric method. Indeed, alcohols such as glycerol,
390
P. Gimeno et al. / Journal of Pharmaceutical and Biomedical Analysis 107 (2015) 386–393
Table 2 Validation data on standard solutions obtained for HPLC–UV method using the ISO 12787. Validation criteria on standard solution (according to ISO 12787) Stability
Standard solutions (H2 O2 and TPP)
>24 h
RSD injection precision TPPO Symmetry (As) Resolution (TPP/TPPO) (Rs)
0.4% 1.04 >1.5
Quantification Limit (LoQ) (based on TPPO peak area)
Value
30 g mL−1 H2 O2
Linearity (H2 O2 + TPP)
Linearity range Determination coefficient R2 Variance Homogeneity (COCHRAN test) Significant slope existence (FISHER) Validity of calibration curve (FISHER) Y-intercept comparison with 0 (STUDENT) Biases
[30–200 g mL−1 ] H2 O2 >0.9995 NS* HS** NS* S*** 1.5 according to the Ph. Eur. 2.2.46) and a number of theoretical plates (N) higher than 10,000 for the TPPO peak. Furthermore, no carryover was pointed out during assay. According to
these results, the chromatographic method can be considered as suitable for the analysis of H2 O2 . Due to the presence of a residual peak of TPPO in TPP standard, a TPPO peak is often present in blank chromatograms involving the necessity to carry out blank preparations for all assays. 3.2.1.2. Stability of standard solutions. The stability of TPP + H2 O2 solutions was checked by re-injecting the six solutions prepared on the first day of the precision study into the HPLC system after 24 h of storage in the HPLC autosampler (dark + 6 ◦ C). The RSD and the confidence interval were determined for both series of values (first injected solutions and re-injected ones). Moreover, the area evolution in 24 h was calculated for each type of solution studied. Results obtained highlight an area increase of the solutions studied in 24 h comprised between 1.5% and 1.8% with an average value of 1.7%. The confidence interval calculated on the areas ranges from 5404 to 5513 with a RSD of 1.2% at T0 and from 5440 to 5603 with a RSD of 1.2% at T0 + 24 h. The overlap of both confidence intervals, respectively calculated at T0 and at T0 + 24 h, and the RSD calculated at 1.5% by grouping the areas obtained after preparation and the areas obtained after 24 h clearly highlight that the area variations observed in 24 h could be considered as negligible. Solutions of TPPO obtained after reaction between H2 O2 and TPP were considered as stable over a period of 24 h if stored in the dark at + 6 ◦ C (i.e. HPLC autosampler). The stability of TPPO was checked and confirmed on prepared cosmetic samples. 3.2.1.3. Quantification limit. The quantification limit (LoQ), corresponding to the lower H2 O2 concentration in a standard solution that could be quantified accurately, was set to 30 g mL−1 corresponding to 0.03–0.07% H2 O2 in cosmetics depending on sample preparation. This limit takes into account the influence of residual TPPO present in TPP standard and allows quantification of H2 O2 at the target value set to 0.1% by the cosmetic regulation in teeth whitening products. This limit is the best one achievable in cosmetics. Depending matrices, this limit could be higher and, if relevant, must be checked according to the ISO 12787 standard (steps 2 and 3). 3.2.1.4. Linearity. Calibration curves were performed three nonconsecutive days using fresh standard solution. Results of the linearity study for the comparison between the curves of
P. Gimeno et al. / Journal of Pharmaceutical and Biomedical Analysis 107 (2015) 386–393 Table 3 Results obtained for the linearity study of hydrogen peroxide and urea peroxide. H2 O2 + TPP
Urea H2 O2 + TPP
Number of calibration level Calibration range Variance homogeneity (COCHRAN test) Aberrant values (DIXON test)
8 [30–200 mg L–1 ] NS*
8 [30–200 mg L–1 ] NS*
No
No
Slope Y-intercept Determination coefficient (R2 ) Concentration deviation of the first standard calibration level (%)
49.44 388.50 0.99973 Less or near 5%
49.72 293.02 0.99957 Less or near 5%
Slope existence (FISHER test) Validity of the calibration curve (FISHER test) Y-intercept comparison with zero (STUDENT test)
HS** NS*
HS** NS*
S***
S***
Slopes comparison Y-intercept comparison * ** ***
NS* NS*
No significant difference (NS). Highly significant difference (HS). Significant difference (S).
TPP + H2 O2 and the curves TPP + urea peroxide using the AVA software are presented in Table 3. The variance homogeneity calculated on calibration values is non-significant (homoscedastic). This allows a valid statistical process of all the linearity criteria. For all curves the determination coefficient (R2 ) is higher than 0.9995 and the bias obtained for the 1st concentration level is lower than 5%. These results display a good linearity correlation for all solutions on the calibration range considered (30–200 g mL−1 ). Y-intercept comparison with zero (STUDENT) is significant in all cases. This might be explained by the fact that commercially available TPP contains approximately 1% TPPO. Comparison between the slopes and Y-intercepts of the curves respectively obtained from hydrogen peroxide solutions and from urea peroxide solutions are non-significant. The two curves can be considered as similar. The results obtained for the determination of hydrogen peroxide and of urea peroxide are comparable. Both, hydrogen peroxide or urea peroxide can be used for standard calibration. 3.2.1.5. Intra- and inter-day precision/residues. Six independent solutions containing each 100 g mL−1 of hydrogen peroxide were prepared everyday during three different days from the corresponding stock solution H2 O2 (prepared every day from commercial H2 O2 ). The intra- and inter-days RSD calculated are both around 1.3% and can thus be considered as acceptable. The nonsignificant variance homogeneity allows a valid statistical process of the intra- and inter-days precision study. The confidence interval calculated on the mean recovery, ranging form 99.5% to 100.9%, is less than 1.5% wide and includes the 100% value. The accuracy of the method can be considered as acceptable. Moreover the fact that the mean validity test is non-significant indicates that there is no day effect. This result meets for example the accuracy requirement for pharmaceutical assay (
